Summary: | Solaris OS compile fix | ||||||
---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Holger Freyther <zecke> | ||||
Component: | New Bugs | Assignee: | Nobody <webkit-unassigned> | ||||
Status: | RESOLVED FIXED | ||||||
Severity: | Normal | CC: | mrowe | ||||
Priority: | P2 | ||||||
Version: | 528+ (Nightly build) | ||||||
Hardware: | PC | ||||||
OS: | OS X 10.4 | ||||||
Attachments: |
|
Description
Holger Freyther
2007-12-23 15:29:41 PST
Created attachment 18080 [details]
Solaris Compile fix from Adrian de Groot
Compile fix for Solaris OS and non GNU compilers.
Comment on attachment 18080 [details]
Solaris Compile fix from Adrian de Groot
Looks fine. strange that: isinf would use one NaN test and isfinite would use a different one. Perhaps we should fix that before landing. !isNan() is clearer.
It would be better if these were written as inline functions, not #defines: +#if PLATFORM(SOLARIS_OS) && COMPILER(GCC) +#include <ieeefp.h> +#if !defined(isfinite) +#define isfinite(x) (finite(x) && !isnand(x)) +#endif +#if !defined(isinf) +#define isinf(x) (!finite(x) && (x == x)) +#endif +#if !defined(signbit) +#define signbit(x) (x < 0.0) +#endif +#endif + Please fix when landing. Does SOLARIS_OS need the _OS suffix? The reason for the _OS suffix on WIN is to distinguish between the base platform and the higher-level APIs. Does Solaris also refer to a higher-level environment to which WebKit may one day be ported? My understanding was that it runs X with GNOME, thus the Gtk or Qt ports would provide this higher layer. Committed revision 29454. I considered all the comments in this bug. The implementation of signbit is definitely incorrect, so I added a FIXME. |